Cantina 发布开源安全研究模型 apex-flash-1，60 个漏洞任务解出 40 个
Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Cantina 把 Z.ai 的 GLM-5.3-Flash 微调成了专挖漏洞的开源模型，60 个私有漏洞任务解出 40 个，MIT 协议随便部署。
Cantina Security 与 Yeta Labs 发布 apex-flash-1，一个专为漏洞挖掘训练的开源权重模型。该模型基于 Z.ai 的 GLM-5.3-Flash 做强化学习微调，在 Hugging Face 以 MIT 许可证开源。在 60 个未公开的漏洞复现任务中，apex-flash-1 解出 40 个。权重可用 vLLM、SGLang 或 Transformers 部署，BF16 精度约需 640 GB GPU 显存。
Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks
Cantina Security, with Yeta Labs, has released apex-flash-1, an open-weights model trained specifically for vulnerability research. It is a reinforcement learning fine-tune of Z.ai’s GLM-5.3-Flash, released on Hugging Face under the MIT license. Is it deployable? Yes, the MIT weights serve on vLLM, SGLang or Transformers, but BF16 needs roughly 640 GB of GPU memory. […] The post Can an Open Model Do Security Research? Cantina’s apex-flash-1 Solves 40 of 60 Held-Out Bug Tasks appeared first on MarkTechPost .
